KYC-Based Voter Verification System Using Facial Authentication and YOLO-Based OCR for ID Card Validation

Abstract

Identity verification of voters is a very important process in the provision of fair and fraud free elections. In many polling stations, voter identity is still verified manually. Election officers compare the voter's face and ID card with the details available in the voter records before allowing them to vote. Although this method has been used for many years, it mainly depends on human observation, which can lead to mistakes. It also increases the possibility of impersonation and the use of fake identity documents. To overcome these issues, this paper presents a KYC-based voter verification system that uses computer vision and deep learning for automatic identity verification. The proposed system performs verification in two stages. First, a face recognition model compares the live image of the voter with the registered image stored in the database. Next, a YOLO-based object detection model, along with Optical Character Recognition (OCR), detects and reads the roll number from the voter's ID card. The extracted roll number is then matched with the database record. A voter is allowed to proceed only when both the face and the ID card details are successfully verified. By combining these two verification methods, the system helps reduce impersonation and provides a more secure and reliable voter verification process.